DOI: 10.18178/wcse.2024.06.027
Developing a Credit Scoring Model for Evaluating Credit Risk in the Nepali Setting
Abstract— Access to credit is vital for economic development, yet Nepal lacks tailored credit scoring models, particularly for retail and micro-lending, due to limited credit history. In this study, we address this gap by developing a locally relevant credit scoring model using data from Foneloan, a digital lending platform. Leveraging machine learning techniques and Python libraries, we identify influential features, determine creditworthiness parameters, and select appropriate algorithms (ADA, DT, ET, RF, XGBoost). Our analysis reveals a significant decrease in bad rate or non-performing loan (NPL) across various score ranges with modified credit scoring model compared to the standard globally used model indicating enhanced loan performance and creditworthiness. This suggests that borrowers with higher credit scores are less likely to default, reflecting better credit management practices. Overall, our study contributes to improving credit risk management in Nepal, fostering economic development and financial inclusion.
Index Terms— Credit Risk Management, Credit Scoring, Digital Lending
Suresh Gautam
Nepal Open University, NEPAL
Bhoj Raj Ghimire
Nepal Open University, NEPAL
Cite: Suresh Gautam, Bhoj Raj Ghimire, "Developing a Credit Scoring Model for Evaluating Credit Risk in the Nepali Setting," 2024 The 14th International Workshop on Computer Science and Engineering (WCSE 2024), pp. 175-181, Phuket Island, Thailand, June 19-21, 2024.